Two Access Methods (AM) waits share this page:
AM_INDBUILD_ALLOCATION: a thread waiting for exclusive access to an extent during allocations while building or rebuilding indexes. Parallel index builds allocate space aggressively, and workers occasionally queue at the same extent.AM_SCHEMAMGR_UNSHARED_CACHE: a thread waiting on the initialisation of a cache of column attributes that cannot be shared between partitions, schema manager plumbing exercised by partitioned objects.
Both are storage engine internals with narrow, well-defined triggers.
Are They a Problem?
No; neither has been a noticeable contention point. AM_INDBUILD_ALLOCATION accrues only during index build activity and scales with build parallelism; treat any accumulation as part of the index maintenance footprint rather than a fault. The schema manager cache wait fires around partitioned-table metadata initialisation and stays in the noise even on heavily partitioned systems.
Slow index builds have better-known bottlenecks (sort memory, tempdb, log throughput) that this allocation wait never headlines.
What To Do
- Filter both in routine analysis.
- For slow index builds, work the standard levers: build MAXDOP, sort memory and tempdb health, off-peak scheduling, and
SORT_IN_TEMPDBplacement.
How To See It
Rank waits with Get-WaitStatistics; expect traces aligned with index maintenance and partitioned-object usage.
These waits are background noise on a healthy instance. If you are chasing a real problem, start with the waits that actually cost you time.
- The waits that usually matter, ranked, with what each one is telling you.
- Performance & Troubleshooting, the scripts and guides for this part of the stack.
Leave a Reply